Model-average estimates (EST) of scaled test predictors of bee biodiversity represented by bee count (BC), Shannon diversity (SD), and species richness (SpR) using bumble bees (bb), solitary bees (sb) and all wild bees (nohb).

The R2 corresponds to the global model including all predictors that remained after model selection, while the R2 of the Null model (R2NULL) refers to the model including only the fixed effects control predictors. Δ = R2- R2NULL. IMP = relative importance; SD = Standard deviation; nmodel = number of models averaged; P = P-values. con2 = 2nd order contrast; ent1 = 1st order entropy; hom2 = 2nd order homogeneity; NDVI_cv = coefficient of variance of the NDVI; rough = surface roughness; 100/1000 = 100 m or 1000 m scale.